See axolotl config

axolotl version: 0.4.1

adapter: qlora
auto_resume_from_checkpoints: true
base_model: Qwen/Qwen1.5-0.5B
bf16: auto
chat_template: llama3
dataset_prepared_path: null
dataset_processes: 6
datasets:
- data_files:
  - ddb87243625694a7_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/ddb87243625694a7_train_data.json
  type:
    field_input: input
    field_instruction: instruction
    field_output: output
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 5
eval_max_new_tokens: 128
eval_steps: 200
eval_table_size: null
evals_per_epoch: null
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: false
hub_model_id: error577/e97aeab0-c073-4258-aff7-d63195974819
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: null
micro_batch_size: 8
mlflow_experiment_name: /tmp/ddb87243625694a7_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 6
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 200
sequence_len: 256
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.005
wandb_entity: null
wandb_mode: online
wandb_name: 574ae888-c1cd-4bb7-b1a1-132d64a5062a
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 574ae888-c1cd-4bb7-b1a1-132d64a5062a
warmup_steps: 30
weight_decay: 0.0
xformers_attention: null

e97aeab0-c073-4258-aff7-d63195974819

This model is a fine-tuned version of Qwen/Qwen1.5-0.5B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.5515

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 30
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Validation Loss
2.2484	0.0006	1	2.4449
1.7868	0.1272	200	1.9947
1.9022	0.2544	400	1.8849
1.7732	0.3816	600	1.8177
1.8026	0.5088	800	1.7686
1.618	0.6360	1000	1.7252
1.7063	0.7632	1200	1.6883
1.4419	0.8904	1400	1.6632
1.5337	1.0176	1600	1.6426
1.5104	1.1449	1800	1.6320
1.695	1.2721	2000	1.6179
1.5855	1.3993	2200	1.6006
1.4552	1.5266	2400	1.5800
1.3088	1.6538	2600	1.5730
1.5115	1.7810	2800	1.5607
1.2866	1.9083	3000	1.5506
1.415	2.0355	3200	1.5562
1.2716	2.1627	3400	1.5532
1.0643	2.2899	3600	1.5532
1.123	2.4171	3800	1.5482
1.186	2.5443	4000	1.5454
1.184	2.6715	4200	1.5438
1.2843	2.7987	4400	1.5430
1.2015	2.9259	4600	1.5426
1.3136	3.0534	4800	1.5699
1.2347	3.1806	5000	1.5621
1.3003	3.3078	5200	1.5594
1.3106	3.4350	5400	1.5544
1.2615	3.5623	5600	1.5515

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1